With the acceleration of globalization, social media has become a core channel for transnational information dissemination, playing an increasingly crucial role in the process of international information "backflow." This study analyzes the impact of social media users' pre-understanding on information polarization behavior through artificial intelligence (AI) recommendation systems and verifies the influence pathways of various factors on polarization behavior using structural equation modeling (SEM). The findings reveal that users' ideological orientations, cultural affiliations, and individual psychological attributes significantly affect their willingness to engage in polarized information behavior. Moreover, narrative comparison serves as a mediating variable between ideological orientation, cultural affiliation, individual psychological attributes, and the willingness to engage in polarization. AI recommendation systems continuously adjust information flows based on users' historical preferences, further exacerbating information polarization. This study provides new perspectives on the cognitive and emotional mechanisms involved in modern information dissemination and offers a theoretical foundation for improving cross-cultural communication and optimizing information technologies.
Read full abstract